cs711: reference monitors part 1: os & sfi greg morrisett cornell university

CS711: Reference MonitorsPart 1: OS & SFI

Greg MorrisettCornell University

June 2001 Lang. Based Security 2

A Reference MonitorObserves the execution of a program and halts

the program if it’s going to violate the security policy.

Common Examples:– operating system (hardware-based)– interpreters (software-based)– firewalls

Claim: majority of today’s enforcement mechanisms are instances of reference monitors.


Reference Monitors Outline• Analysis of the power and limitations.

• What is a security policy?• What policies can reference monitors enforce?

• Traditional Operating Systems.– Policies and practical issues– Hardware-enforcement of OS policies.

• Software-enforcement of OS policies.– Why?– Software-Based Fault Isolation – Java and CLR Stack Inspection

– Inlined Reference Monitors


Requirements for a Monitor• Must have (reliable) access to information about what

the program is about to do.– e.g., what instruction is it about to execute?

• Must have the ability to “stop” the program– can’t stop a program running on another machine that you

don’t own.– really, stopping isn’t necessary, but transition to a “good”

state.

• Must protect the monitor’s state and code from tampering.– key reason why a kernel’s data structures and code aren’t

accessible by user code.

• In practice, must have low overhead.


What Policies?We’ll see that under quite liberal assumptions:

– there’s a nice class of policies that reference monitors can enforce (safety properties).

– there are desirable policies that no reference monitor can enforce precisely.

• rejects a program if and only if it violates the policy

Assumptions:– monitor can have access to entire state of

computation.– monitor can have infinite state.– but monitor can’t guess the future – the predicate it

uses to determine whether to halt a program must be computable.


Schneider's FormalismA reference monitor only sees one execution

sequence of a program.So we can only enforce policies P s.t.:

(1) P(S) = S.P ()

where P is a predicate on individual sequences.

A set of execution sequences S is a property if membership is determined solely by the sequence and not the other members in the set.


More Constraints on MonitorsShouldn’t be able to “see” the future.

– Assumption: must make decisions in finite time.– Suppose P () is true but P ([..i]) is false for

some prefix [..i] of . When the monitor sees [..i] it can’t tell whether or not the execution will yield or some other sequence, so the best it can do is rule out all sequences involving [..i] including .

So in some sense, P must be continuous:

(2) .P () (i.P([..i]))


Safety PropertiesA predicate P on sets of sequences s.t.

(1) P(S) = S.P ()

(2) .P () (i.P([..i]))

is a safety property: “no bad thing will happen.”

Conclusion: a reference monitor can’t enforce a policy P unless it’s a safety property. In fact, Schneider shows that reference monitors can (in theory) implement any safety property.


Safety vs. SecuritySafety is what we can implement, but is it what

we want?– “lack of info. flow” isn’t a property.

Safety ensures something bad won’t happen, but it doesn’t ensure something good will eventually happen:– program will terminate– program will eventually release the lock– user will eventually make payment

These are examples of liveness properties. – policies involving availability aren’t safety prop. – so a ref. monitor can’t handle denial-of-service?


Safety Is NiceSafety does have its benefits:

– They compose: if P and Q are safety properties, then P & Q is a safety property (just the intersection of allowed traces.)

– Safety properties can approximate liveness by setting limits. e.g., we can determine that a program terminates within k steps.

– We can also approximate many other security policies (e.g., info. flow) by simply choosing a stronger safety property.


Practical IssuesIn theory, a monitor could:

– examine the entire history and the entire machine state to decide whether or not to allow a transition.

– perform an arbitrary computation to decide whether or not to allow a transition.

In practice, most systems:– keep a small piece of state to track history– only look at labels on the transitions– have small labels– perform simple tests

Otherwise, the overheads would be overwhelming.– so policies are practically limited by the vocabulary of labels,

the complexity of the tests, and the state maintained by the monitor.


Reference Monitors OutlineAnalysis of the power and limitations.

What is a security policy?What policies can reference monitors enforce?

• Traditional Operating Systems.– Policies and practical issues– Hardware-enforcement of OS policies.

• Software-enforcement of OS policies.– Why?– Software-Based Fault Isolation– Inlined Reference Monitors


Operating Systems circa ‘75Simple Model: system is a collection of running

processes and files.– processes perform actions on behalf of a user.

• open, read, write files• read, write, execute memory, etc.

– files have access control lists dictating which users can read/write/execute/etc. the file.

(Some) High-Level Policy Goals:– Integrity: one user’s processes shouldn’t be able to

corrupt the code, data, or files of another user. – Availability: processes should eventually gain

access to resources such as the CPU or disk.– Secrecy? Confidentiality? Access control?


What Can go Wrong?– read/write/execute or change ACL of a file for

which process doesn’t have proper access.• check file access against ACL

– process writes into memory of another process• isolate memory of each process (& the OS!)

– process pretends it is the OS and execute its code• maintain process ID and keep certain operations

privileged --- need some way to transition.

– process never gives up the CPU• force process to yield in some finite time

– process uses up all the memory or disk• enforce quotas

– OS or hardware is buggy...


Key Mechanisms in Hardware– Translation Lookaside Buffer (TLB)

• provides an inexpensive check for each memory access.• maps virtual address to physical address

– small, fully associative cache (8-10 entries)

– cache miss triggers a trap (see below)

– granularity of map is a page (4-8KB)

– Distinct user and supervisor modes• certain operations (e.g., reload TLB, device access)

require supervisor bit is set.

– Invalid operations cause a trap• set supervisor bit and transfer control to OS routine.

– Timer triggers a trap for preemption.


Steps in a System CallTime

calls f=fopen(“foo”)

User Process

library executes “break”

Kernel

trapsaves context, flushes TLB, etc.

checks UID against ACL, sets up IO buffers & file context, pushes ptr to context on user’s stack, etc.

restores context, clears supervisor bitcalls fread(f,n,&buf)library executes “break”

saves context, flushes TLB, etc.checks f is a valid file context, doesdisk access into local buffer, copiesresults into user’s buffer, etc.

restores context, clears supervisor bit


Hardware TrendsThe functionality provided by the hardware hasn’t

changed much over the years. Clearly, the raw performance in terms of throughput has.

Certain trends are clear:– small => large # of registers: 8 16-bit =>128 64-bit– small => large pages: 4 KB => 16 KB– flushing TLB, caches is increasingly expensive– computed jumps are increasingly expensive– copying data to/from memory is increasingly expensive

So a trap into a kernel is costing more over time.


OS TrendsIn the 1980’s, a big push for microkernels:

– Mach, Spring, etc.– Only put the bare minimum into the kernel.

• context switching code, TLB management• trap and interrupt handling• device access

– Run everything else as a process.• file system(s)• networking protocols• page replacement algorithm

– Sub-systems communicate via remote procedure call (RPC)

– Reasons: Increase Flexibility, Minimize the TCB


A System Call in MachTime

f=fopen(“foo”)

User Process

“break”

Kernel

saves contextchecks capabilities,copies argumentsswitches to Unixserver context

Unix Server

checks ACL, sets upbuffers, etc.

“returns” to user. saves contextchecks capabilities,copies results

restores user’scontext


MicrokernelsClaim was that flexibility and increased assurance

would win out. – But performance overheads were non-trivial – Many PhD’s on minimizing overheads of communication– Even highly optimized implementations of RPC cost 2-3

orders of magnitude more than a procedure call.

Result: a backlash against the approach.– Windows, Linux, Solaris continue the monolithic tradition.

• and continue to grow for performance reasons (e.g., GUI) and for functionality gains (e.g., specialized file systems.)

– Mac OS X, some embedded or specialized kernels (e.g., Exokernel) are exceptions. VMware achieves multiple personalities but has monolithic personalities sitting on top.


Performance MattersThe hit of crossing the kernel boundary:

– Original Apache forked a process to run each CGI: • could attenuate file access for sub-process

• protected memory/data of server from rogue script

• i.e., closer to least privilege

– Too expensive for a small script: fork, exec, copy data to/from the server, etc.

– So current push is to run the scripts in the server. • i.e., throw out least privilege

Similar situation with databases, web browsers, file systems, etc.


The Big Question?From a least privilege perspective, many

systems should be decomposed into separate processes. But if the overheads of communication (i.e., traps, copying, flushing TLB) are too great, programmers won’t do it.

Can we achieve isolation and cheap communication?


Reference Monitors OutlineAnalysis of the power and limitations.

What is a security policy?What policies can reference monitors enforce?

Traditional Operating Systems.Policies and practical issuesHardware-enforcement of OS policies.

• Software-enforcement of OS policies.Why?– Software-Based Fault Isolation– Java Stack Inspection– Inlined Reference Monitors


Software Fault Isolation (SFI)• Wahbe et al. (SOSP’93)• Keep software components in same hardware-based

address space.• Use a software-based reference monitor to isolate

components into logical address spaces.– conceptually: check each read, write, & jump to make sure

it’s within the component’s logical address space.– hope: communication as cheap as procedure call.– worry: overheads of checking will swamp the benefits of

communication.

• Note: doesn’t deal with other policy issues– e.g., availability of CPU


One Way to SFIvoid interp(int pc, reg[], mem[], code[], memsz, codesz) {

while (true) { if (pc >= codesz) exit(1); int inst = code[pc], rd = RD(inst), rs1 = RS1(inst), rs2 = RS2(inst), immed = IMMED(inst);

switch (opcode(inst)) { case ADD: reg[rd] = reg[rs1] + reg[rs2]; break; case LD: int addr = reg[rs1] + immed; if (addr >= memsz) exit(1); reg[rd] = mem[addr]; break; case JMP: pc = reg[rd]; continue;

... } pc++;}}

0: add r1,r2,r3

1: ld r4,r3(12)

2: jmp r4


Pros & Cons of InterpreterPros:

– easy to implement (small TCB.)– works with binaries (high-level language-

independent.)– easy to enforce other aspects of OS policy

Cons:– terribly execution overhead (x25? x70?)

but it’s a start.


Partial Evaluation (PE)A technique for speeding up interpreters.

– we know what the code is.– specialize the interpreter to the code.

• unroll the loop – one copy for each instruction• specialize the switch to the instruction• compile the resulting code

For a cool example of this, see Fred Smith's thesis (hanging off my web page.)


Example PE

Specialized interpreter: reg[1] = reg[2] + reg[3];

addr = reg[3] + 12;

if (addr >= memsz) exit(1);

reg[4] = mem[addr];

pc = reg[4]

0: add r1,r2,r3

1: ld r4,r3(12)

2: jmp r4

...

Original Binary:while (true) { if (pc >= codesz) exit(1); int inst = code[pc]; ...}

Interpreter

0: add r1,r2,r3

1: addi r5,r3,12

2: subi r6,r5,memsz

3: jab _exit

4: ld r4,r5(0)

...

Resulting Compiled Code


SFI in PracticeUsed a hand-written specializer or rewriter.

– Code and data for a domain in one contiguous segment.• upper bits are all the same and form a segment id.• separate code space to ensure code is not modified.

– Inserts code to ensure stores [optionally loads] are in the logical address space.

• force the upper bits in the address to be the segment id• no branch penalty – just mask the address• may have to re-allocate registers and adjust PC-relative offsets

in code.• simple analysis used to eliminate unnecessary masks

– Inserts code to ensure jump is to a valid target• must be in the code segment for the domain• must be the beginning of the translation of a source instruction• in practice, limited to instructions with labels.


More on Jumps• PC-relative jumps are easy:

– just adjust to the new instruction’s offset.

• Computed jumps are not:– must ensure code doesn’t jump into or around a

check or else that it’s safe for code to do the jump.– for this paper, they ensured the latter:

• a dedicated register is used to hold the address that’s going to be written – so all writes are done using this register.

• only inserted code changes this value, and it’s always changed (atomically) with a value that’s in the data segment.

• so at all times, the address is “valid” for writing.• works with little overhead for almost all computed jumps.


More SFI DetailsProtection vs. Sandboxing:

– Protection is fail-stop:• stronger security guarantees (e.g., reads)• required 5 dedicated registers, 4 instruction sequence• 20% overhead on 1993 RISC machines

– Sandboxing covers only stores• requires only 2 registers, 2 instruction sequence• 5% overhead

Remote Procedure Call:– 10x cost of a procedure call– 10x faster than a really good OS RPC

Sequoia DB benchmarks: 2-7% overhead for SFI compared to 18-40% overhead for OS.


Questions• What happens on the x86?

– small # of registers– variable-length instruction encoding

• What happens with discontiguous hunks of memory?

• What would happen if we really didn’t trust the extension?– i.e., check the arguments to an RPC?– timeouts on upcalls?

• Does this really scale to secure systems?

cs711: reference monitors part 1: os & sfi greg morrisett cornell university

Documents

policies p

policy p

predicate p

monitors state

based security9 safety

based security7

based security2

based security5